Estimating Long Term Evolution Network Capacity and Performance

1. A method comprising: receiving, via a computing device, a selection of a target network resource for evaluation; determining, via the computing device, key performance indicators for computing a target network resource value for the target network resource, wherein the determined key performance indicators are filtered out of a set of key performance indicators using statistical criteria and statistical approaches; gathering a data set, the data set comprising a plurality of points, wherein each point indicates a key performance indicator value for at least one of the determined key performance indicators and the target network resource value; training, via incremental training, a calculation module to calculate the selected target network resource value based on key performance indicator values included in the gathered data set, wherein the incremental training comprises updating the calculation module after each additional key performance indicator value set is presented to the calculation module; testing the calculation module using a testing data set, the testing data set having a same probability distribution of key performance indicator values and target network resource values as the gathered data set; and providing, as output, an indication that the calculation module has been trained.

2. The method of claim 1 , further comprising: calculating, using the trained calculation module, a current value of the selected target network resource; determining that the current value of the selected target network resource is not optimal; and providing, as output, an indication of network infrastructure or network components to be added to the network for making optimal the current value of the selected target network resource.

3. The method of claim 1 , further comprising: determining, for each key performance indicator value, at least one of: a trend component, the trend component representing a long-term trend of the key performance indicator value, wherein the long-term trend is a monthly trend or a yearly trend; a seasonality component, the seasonality component corresponding to a periodicity of the key performance indicator value, wherein the periodicity of the key performance indicator value has a period of seven days; a burst component, the burst component representing a predictable outlier of the key performance indicator from a combination of the trend component and the seasonality component; or a random error component, the random error component representing random error unaccounted for by the trend component, the seasonality component, and the burst component; and modifying the calculation module to calculate the selected target network resource value based on time using the trend component, the seasonality component, the burst component, or the random error component.

4. The method of claim 3 , further comprising: determining that the key performance indicator value is computed according to an equation: KPI=(1+B)*(T+S+R), wherein: KPI represents the key performance indicator value, B represents the burst component, T represents the trend component, S represents the seasonality component, and R represents the random error component.

5. The method of claim 1 , wherein, for at least a first key performance indicator value of a first key performance indicator, the calculation module calculates the target network resource value according to an equation: NR=A*e −B*KPI +C, wherein: NR represents the target network resource value, e represents 2.71828, and A, B, and C are constants.

6. The method of claim 5 , wherein A, B, or C are determined based on other key performance indicator values for key performance indicators different from the first key performance indicator.

7. The method of claim 1 , further comprising predicting, using the trained calculation module, a new target network resource value corresponding to new key performance indicator values.

8. The method of claim 1 , wherein the target network resource comprises one or more of an average number of connected users per cell, a downlink physical resource blocks (PRB) usage, an uplink physical resource blocks (PRB) usage, a physical downlink control channel (PDCCH) usage, a paging resource usage, an average number of active users in an uplink buffer per cell, an average number of active users in a downlink buffer per cell, or a physical random access channel (PRACH) usage.

9. The method of claim 1 , wherein the set of key performance indicators comprises one or more of a voice call drop rate, a data call drop rate, a radio resource control (RRC) connection success rate, or a radio access bearer (RAB) setup success rate.

10. The method of claim 1 , wherein the statistical criteria comprises one or more of Akaike Information Criteria (AIC), Bayesian Information Criteria, Minimum Redundancy Maximum Relevance (MRMR), or Lasso.

11. The method of claim 1 , wherein the statistical approaches comprises one or more of improving model interpretability, shortening a model training time, enhancing generalization by reducing over-fitting in the model, or maximizing mean accuracy rate and minimizing mean error rate.

12. A non-transitory computer-readable medium comprising instructions which, when executed by a computer, cause the computer to: receive a selection of a target network resource for evaluation; determine key performance indicators for computing a target network resource value for the target network resource, wherein the determined key performance indicators are filtered out of a set of key performance indicators using statistical criteria and statistical approaches; gather a data set, the data set comprising a plurality of points, wherein each point indicates a key performance indicator value for at least one of the determined key performance indicators and the target network resource value; train, via incremental training, a calculation module to calculate the selected target network resource value based on key performance indicator values included in the gathered data set, wherein the incremental training comprises updating the calculation module after each additional key performance indicator value set is presented to the calculation module; test the calculation module using a testing data set, the testing data set having a same probability distribution of key performance indicator values and target network resource values as the gathered data set; and provide, as output, an indication that the calculation module has been trained.

13. The computer-readable medium of claim 12 , further comprising instructions which, when executed by the computer, cause the computer to: determine, for each key performance indicator value, at least one of: a trend component, the trend component representing a long-term trend of the key performance indicator value, wherein the long-term trend is a monthly trend or a yearly trend; a seasonality component, the seasonality component corresponding to a periodicity of the key performance indicator value, wherein the periodicity of the key performance indicator value has a period of seven days; a burst component, the burst component representing a predictable outlier of the key performance indicator from a combination of the trend component and the seasonality component; or a random error component, the random error component representing random error unaccounted for by the trend component, the seasonality component, and the burst component; and modify the calculation module to calculate the selected target network resource value based on time using the trend component, the seasonality component, the burst component, or the random error component.

14. The computer-readable medium of claim 13 , further comprising instructions which, when executed by the computer, cause the computer to: determine that the key performance indicator value is computed according to an equation: KPI=(1+B)*(T+S+R), wherein: KPI represents the key performance indicator value, B represents the burst component, T represents the trend component, S represents the seasonality component, and R represents the random error component.

15. The computer-readable medium of claim 12 , wherein, for at least a first key performance indicator value of a first key performance indicator, the calculation module calculates the target network resource value according to an equation: NR=A*e −B*KPI +C, wherein: NR represents the target network resource value, e represents 2.71828, and A, B, and C are constants.

16. The computer-readable medium of claim 15 , wherein A, B, or C are determined based on other key performance indicator values for key performance indicators different from the first key performance indicator.

17. The computer-readable medium of claim 12 , further comprising instructions which, when executed by the computer, cause the computer to predict, using the trained calculation module, a new target network resource value corresponding to new key performance indicator values.

18. The computer-readable medium of claim 12 , wherein the target network resource comprises one or more of an average number of connected users per cell, a downlink physical resource blocks (PRB) usage, an uplink physical resource blocks (PRB) usage, a physical downlink control channel (PDCCH) usage, a paging resource usage, an average number of active users in an uplink buffer per cell, an average number of active users in a downlink buffer per cell, or a physical random access channel (PRACH) usage.

19. The computer-readable medium of claim 12 , wherein the set of key performance indicators comprises one or more of a voice call drop rate, a data call drop rate, a radio resource control (RRC) connection success rate, or a radio access bearer (RAB) setup success rate.

20. A system comprising: one or more processors; and a memory storing instructions which, when executed by the one or more processors, cause the one or more processors to: receive a selection of a target network resource for evaluation; determine key performance indicators for computing a target network resource value for the target network resource, wherein the determined key performance indicators are filtered out of a set of key performance indicators using statistical criteria and statistical approaches; gather a data set, the data set comprising a plurality of points, wherein each point indicates a key performance indicator value for at least one of the determined key performance indicators and the target network resource value; train, via incremental training, a calculation module to calculate the selected target network resource value based on key performance indicator values included in the gathered data set, wherein the incremental training comprises updating the calculation module after each additional key performance indicator value set is presented to the calculation module; test the calculation module using a testing data set, the testing data set having a same probability distribution of key performance indicator values and target network resource values as the gathered data set; and provide, as output, an indication that the calculation module has been trained.